SMX East 2017 Day 2 - Paid search trends, Amazon and Automation

SMX East 2017 Day 2 – Paid search trends, Amazon and Automation

Day 2 of #SMX East offers new insights on paid search trends, what Amazon is doing in search engine marketing, and SEM automation.

SMX East 2017, Day 2
QuanticMind is back for day 2 of SMX East 2017, with more insights and best practices from industry experts. Here’s our list of the quick hits:

>> Quick Links to Day 2 Session Recaps

Day 2 Keynote – Search Marketers Are The Trailblazers In The Emerging Martech Era
Featuring: @chiefmartec

  • Search marketers will, according to Brinker, “inherit the earth, or at least martech.
  • The originator of the infamous martech infographics reviews how the original versions from 2011 listed several hundred logos, while the 2014 version had 1,000 logos of distinct marketing technology companies; the 2015 version had close to 2,000; and the 2017 version has more than 5,000, despite the constant assurances of pundits and media that, like with the “death” of SEO, the martech space would consolidate.
  • Brinker points out, however, that there has been some consolidation in that smaller marketing firms have been getting acquired by larger firms – and the original founders of these smaller firms then cash out after acquisition and go off to fund dozens more smaller firms. These smaller firms are also driving significant marketing innovation.
  • In order to appreciate the martech landscape, it’s useful to consider certain key shifts that have changed in recent years:
    • Lower barrier of entry for software development – It’s never been easier to build software now that so many assets are available in an open-source format (rather than having to buy expensive development licenses). Additional developments such as cloud storage and sharing, free real-time chatting and conference communications and the increased feasibility of distributed teams across the world has made developing software exponentially easier than it was 10 years ago.
    • Lower barrier to selling software – Software as a service (SaaS) and freemium (initial free trial with the option to upgrade to a paid subscription) are have become extremely popular business models with strong adoption. Studies show that on average, as of April of this year, enterprise companies in a variety of verticals used, on average anywhere from 800-1200 cloud-based software applications. Of these, marketing technologies were the most commonly purchased (though other popular software categories included project management/collaboration, human resources management and other types).
    • Entrepreneurship is growing across multiple verticals – Factors such as the viability of smaller and virtual businesses, consumer tech disruptions such as mobile phones, globalization and Moore’s Law have helped marketing, in Brinker’s words, “move beyond advertising.”
    • Greater appetite for marketing technology – On the buy site, the appetite for software is enormous, thanks to the such changes as th growth of new marketing channels, more strategic departmental intersections, a growing emphasis on full customer lifecycle tracking and the growing role of marketing as a revenue producer.
  • Brinker notes that for all these significant touchpoints, marketers increasingly act as the central hub that owns these activities, similar to the way IT departments were once popularly the gatekeepers of all enterprise technology in the 1990s and early 2000s. “[Marketers] don’t run everything that touches everything,” says Brinker,  “But we provide governance for how technology interacts with customers.”
  • Brinker also notes that in many cases, marketing technology has come to expand in a very strategic manner with minimal overlap. Companies he classifies in the digital engagement tech stack, for instance, such as Demandbase, AppAnnie and BlueKai evince lots of capability and specialization with very little functional overlap. This means that the great expansion of martech may not necessarily be characterized by crowded markets of me-too applications, as they might have been during historical boom periods.
  • Brinker notes that martech sits at the nexus of marketing, technology and management, and that major media publications such as Harvard Business Review, Financial Times and Forbes have all started to embrace the concept of a “Chief Martech Officer.” In a recent survey, 53% of surveyed companies stated they already have a staffer explicitly in charge of martech, and 21% planned to add someone in the role in the near future.
  • Independent analyst firm TBR reports that 75%+ surveyed companies have an internal martech executive. The responsibilities of such executives include researching and recommending new martech solutions, bringing martech into business units to drive adoption, and providing training.
  • Another study showed that approximately 60% of all included organizations plan to restructure their marketing stack in the next 12 months to take advantage of new martech solutions.
  • Brinker caps his session with a list of “9 reasons why search marketers make great martech leaders”:
  1.  Familiarity with managing an evolving tech landscape – including SEO, content, social, mobile, video, call, local, affiliate and e-commerce.
  2. Tendency to approach performance-based marketing analytically
  3. Tendency to embrace A/B testing to bridge the gap between “marketing art and science”
  4. Familiarity with running agile marketing and testing to quickly iterate with feedback loops
  5. Ability to craft effective user experience, such as by way of conversion optimization and tracking the customer journey.
  6.  Ability to operate digital marketing campaigns at massive scale
  7. Ability to adapt to major environmental changes, such as changes from Google to algorithms and bidding
  8. Ability to strategically blend paid, owned and earned campaigns
  9. Ability to design contextual customer journeys across multiple touchpoints, from search queries to ad copy to landing pages to follow-up.


>>Back to top

What Is Customer First Advertising? Why Should I Care?
Featuring: Colleen Aubrey of @amazon

  • Amazon has made its SMX debut to discuss how to “how to work on the Shopping experience, working backward to advertisers, and create value for all audiences: shoppers, advertisers and publishers.” (In this case, Amazon is the only publisher in play.)
  • Amazon’s goal with its ad program is to drive discovery within the Amazon experience, focusing on products sold on Amazon. The primary goals that the company envisions for advertising on Amazon are to drive brand engagement and downstream sales volume.
  • Aubrey notes that this philosophy stems from Amazon’s “customer-obsessed” approach of focusing on customers first. This translates into the company’s goal of, first and foremost, creating a great Shopping experience, where ads are so unobtrusive and have such relevant placements that they aren’t recognized as being ads. This should create value for advertisers as shoppers become more engaged and buy more frequently, and is ultimately good for the publisher (again, Amazon, in this case, though the company also has a programmatic business with multiple publisher customers online).
  • Amazon’s focus is on price, selection and convenience, which has led its customer team to develop a large, engaged customer base. Aubrey notes that a significant number of Amazon customers begin their Shopping experience with the Amazon search bar, and use it to generate a “specific signal,” which the company considers to be different from generalized search engines and more specifically to shopping for specific products. The company uses this signal to try to understand shopper and brand intent.
  • The company has, of course, made numerous forays into digital advertising, with mixed results. Its Amazon Product Ads, a comparison shopping product introduced some years ago, were considered to be very disruptive to search advertisers. The company also ran a text-based ad service known as ClickRiver to drive external traffic to hard-to-sell products. Both products have been discontinued.
  • In her experience working on Product Ads, Aubrey and her team were tasked with ensuring ad content was accurate and properly linked, but noticed that advertisers were creating campaigns pointing not to specific product detail listings, but rather to category pages – a practice that had previously not been allowed. They did so because advertisers “wanted to drive experience for their products.” This has helped inform the company’s current digital advertising strategy and ad units:
    • Headline Search Ads (HSA) – These image-based ads are brand-level units that show categories of items. A demonstration search for “denim tote” pulled up the HSA for a specific handbag company whose ad featured a prominent handbag image and copy specifying “handbags, backpacks, totes, satchels and more.” Clicking on the first unit in this ad usually takes shoppers to either a multi-product landing page or full-blown store page, while clicking on the subsequent units goes directly to a specific product listing for that merchant. (It should be noted that items that are out of stock will not be listed in the HSA, and if all designated HSA products are out of stock, that HSA simple will not run on Amazon.) Amazon hopes these store pages will serve the dual needs of providing vendors with a virtual storefront while also providing traffic of interested buyers to make a match. These storefronts let vendors curate product collections and even surface an “About Us” story about the vendor company itself.
    • Sponsored Products Ads – These ad units appear in multiple placements on Amazon, including related product streams, product detail pages, thank you pages and post-checkout pages.
    • Product Display Ads – These are product-level ads that are targeted by keywords but do not allow any custom creative – all images and copy are sourced directly from product details themselves. Amazon endeavors to have these ads appear only in highly relevant areas to avoid disrupting the shopping experience. These are CPC ads that can also be used on product detail pages to cross-sell or upsell related products.
  • Ad strategy – While advertisers can and do place their product ads onto competitor pages, the Amazon team has seen several other use cases, such as merchants running ads for complimentary or accessory products, or using a “good-better-best” strategy to display ads for higher-priced or lower-priced products in the same product line.
  • Amazon ad ranking and lack of Quality Score – Amazon’s primary objective with digital ads is to find a good fit between relevant ads and shoppers that are in-market for that advertiser’s products. However, if ad campaigns show a low engagement rate, Amazon’s current system will flag those ads as being of lower quality and rank them lower, such that they surface less. There is no actual Amazon version of Google’s Quality Score, but the Amazon team is looking into this, and other ways to improve its ad algorithms. There is no clear timeframe within which an Amazon ad may get marked as being a low-quality ad; the main drivers of the ranking are impressions to clicks to conversions, and to a lesser extent, bid levels.
  • Ad best practices – Unfortunately, Amazon’s digital advertising best practices are something of a black box. The publisher has not displayed its advertising best practices publicly. This is something the Amazon ad team is “looking at,” but may not “have the right solution yet.”
  • Amazon’s Choice, Prime designations – It should be noted that the “Amazon’s Choice product designation is completely customer-chosen, and is generated by customer ratings and experiences. It can’t be influenced by vendors or advertisers. Products can be designated as Prime either when they are sold by vendors who partner directly with Amazon to manage inventory and fulfillment, or when they are able to reliably fulfill 2-day shipping orders on their own.
  • 5 factors for success with Amazon digital ads – Aubrey recommends these 5 tips:
    • Start with Sponsored Products Ad units – Assemble a group of products, recommended keywords and consider using Amazon’s suggested bid levels and using the publisher’s auto-targeting. The system currently does not cater to small-scale testing of one item with a minuscule bid.
    • Use Headline Search and Product Display Ads for more use cases – Amazon recommends that once advertisers build up familiarity with listings, they then explore HAS units, then move on to Product Display Ads, which will offer lower traffic but potentially higher specificity.
    • Enable “Bid+” – Sponsored Product Ads have a built-in bidding option that automatically increases your bids to up to 50% more than your standard budget limit to increase the chances of getting placement during high-traffic areas. This is a cost-per-click option that will not overrun budget if no clicks come in.
    • Opt for always-on ads – Ads can be set to having a duration of “No End Date” at any time. Since they run on a CPC basis, there is no risk of wasting spend if there are simply no clicks at all, only an opportunity cost if, for instance, items that vendors are selling get clicks during off-season periods.
    • Build an Amazon storefront – Aubrey recommends eventually getting into the habit of providing a curated shopping experience to help communicate your brand’s identity and build stronger engagement with current and potential shoppers.
  • Additional e-commerce tips:
    • Carefully vet product detail pages – Apply a critical eye to your product detail pages to make sure they’re set up for success, ensuring they are information rich regarding specifications on sizes, colors, materials and other important purchase info. Product detail pages that are “thin” on information or have poor customer reviews will undermine ad performance – at this point, a better strategy would be to shut off ads for such poor performers entirely, focusing instead on your strongest SKUs.
    • Keep products in stock – On Amazon, as mentioned, product-level ads won’t be served if they are not in stock. Headline ads will still be served but may be populated with your less-popular items, leading to worse performance if you’re out of inventory for your bestsellers.

>>Back to top

Artificial Intelligence Ain’t Art
Featuring: @MerryMorud

  • Morrud: “None of you [PPC professionals] should be concerned about being automated out of a job.”
  • Creative is the real differentiator of brands, but SEM professionals arguably think about it least, since they often focus too much on operations.
  • Artificial intelligence (AI) still needs human input for matters such as business intelligence, ingenuity and creative.
  • Computers can be programmed with all the possible moves and outcomes in a chess game, but can’t understand that humans are inherently irrational, and will occasionally exhibit wildly unpredictable behavior.
  • AIs can’t set KPIs, nor can they ask the right questions? For instance, when attempting to target prospective customers for an expensive enterprise SaaS product, “Facebook will tell you to target everybody” and let the algorithm sort it out. However, for such a specialized and big-ticket product, your audience isn’t everybody.
  • Algorithms require large data sets to be most effective; a big audience to optimize against. Advertisers with small, niche industries, they may not be well-served by algorithms.
  • As an analogy, human creativity could be compared to the musical genius of Freddie Mercury and his band Queen which creates entirely new sounds, while AI is, at best, a cover band, iterating on what has already been created.
  • It’s been suggested that up to 75% of the likelihood of your ad being seen by a user has to do with creative – that it’s not enough to toss out the same old CTA. Innovative campaigns such as Dove’s “Real Beauty” campaign or AirBNB’s “Belong Anywhere” campaign could not have been created by an AI.
  • Case studies: Morrud shares examples in which real-life events, such as sporting events and high-profile TV premieres drove higher campaign performance her team had observed from automation.
  • At AdWeek 2017, diversity was a big topic – Teen Vogue grew 500% taking on taboo subjects regarding social issues, while an AdWeek study suggested that “ads that are culturally progressive perform 25% better” than those that aren’t. This is borne out in practice by algorithms that might focus on affluent 18-34-year-old caucasian shoppers while ignoring lucrative audiences from other ethnic groups.

>>Back to top

Of Machines and Men
Featuring: @AnnieCushing

  • AI as we know it could be defined as an intelligence layer which, when placed atop a huge mound of data, can then analyze that data much faster than humans can.
  • Some real-world examples of how AI is helping people today:
    • FarmLogs – A farm management mobile app that uses AI to monitor land, weather and soil conditions to give practical recommendations on finding arable land, given that, by 2050, the earth’s population is expected to reach 9.7B people, requiring a 50% increase in farmable land.
    • Intel Precision Medicine – Spearheaded by an Intel marketer who is also a cancer patient, this initiative is working on an AI which, by 2020, is expected to be able to diagnose a personalized cancer treatment plan within one day based on a patient’s DNA and medical history.
    • Predictive suicide prevention – A Florida State University scientist has worked out an AI which is being used to recognize suicide risk factors within 2 years with an accuracy rating of 80-90%.
    • Predictive terrorism prevention – Facebook is now using algorithms to automatically detect images and language patterns that might indicate terrorist activity.
  • There are also several examples of AI being used outside of sciences and academia, such as chatbots, voice assistants and even predictive shopping algorithms on For instance, some brands are using AIs to track brand mentions and assess associated sentiment – positive or negative.
  • While there are tragic examples of AI gone wrong, such as a self-driving Tesla vehicle that struck and killed an infant, there are also bizarre examples of AI simply being not up to part with human cognition:
    • AI-judged international beauty pageant – Last year, the first international beauty contest judged by an AI took place. Unfortunately, the AI wasn’t properly trained to recognize any standards of beauty beyond caucasian women, so its choice of winners was embarrassingly non-diverse.
    • Fast-food purveyor Burger King ran an ad asking Google Now the ingredients of its famous Whopper hamburger sandwich. Google Now responded by reading out the contents of the Whopper page on Wikipedia. However, since Wikipedia is editable by the general public, Google Now quickly started stating that Whoppers contain cyanide, rats and small children, causing Google to intervene and shut down the query.
    • Sophia at SXSW – Most recently, the robot Sophia from Hanson Robotics was featured at the South By Southwest event, interviewed by a human interviewer who jokingly asked the talking robot about destroying the human race. The robot immediately agreed to destroy the human race.
  • Future opportunities to use AI – Cushing sees additional, day-to-day use cases for AI, such as for predictively prescribing travel tips for passengers with pre-booked flights, or having an Apple Watch or similar exercise monitor intelligently shut off heart rate monitors after the owner’s heart rate drops or moves away from the gym to a different location measured by GPS.

>>Back to top

The Robots are Coming?!
Featuring Elizabeth Clark of @dreamagility

  • Marketers can certainly “do clever stuff with AI,” but AI still needs human input. A data scientist recently added the content of 20,000 novels into a neural net and requested it generate the ideal romance novel title. It returned the incoherent “Lady Convenience Wife.”
  • Clark recounts starting her new business from scratch without the aid of a huge team of assistants. To drive more business to her website, which focused on dating and flirting, she hit on the brilliant idea of sending out a marketing email with what were then the three most popular words online: “free,” “sex” and “chocolate.” This led to an absolutely massive open rate. Unfortunately, this was only a one-time success.
  • Some notable figures about the rise of marketing automation:
    • +451% lift in qualified leads
    • +85% revenue growth and +69% sales productivity through lead nurturing
    • 49% of surveyed companies use some form of marketing automation
    • 63% of surveyed companies using marketing automation are outpacing competitors that don’t use it
    • B2B marketers using automation increased sales pipeline contribution 10%
    • 36% of marketers use automation to delegate repetitive tasks
  • In essence, the benefits of marketing automation are that they lower costs, free up time, have less room for human error, and can drive higher output.
  • Still, human input remains vital, especially to tap into what Clark calls the “five big emotions” that drive marketing and conversion rates: Pride, trust, anticipation, fear, and joy.

>>Back to top

Marketing Partners: Essential Allies or Necessary Frenemies?
Featuring: @jknob8

  • Collins explains that as VP of Integrated Brand Marketing at financial giant Morgan Stanley, her team is less interested in brand building and more interested in lead generation for a network of 15,000 local financial advisors – essentially franchisees – working out of regional Morgan Stanley offices.
  • While these internal partners are financially savvy, they occasionally proved to be less than adept at search engine marketing, using copy, images and promotional tactics that were off-brand and off-message for Morgan Stanley’s standard digital programs.
  • In addition, Collins saw these partners become “frenemies” to her team’s budget as they began aggressively bidding for national keywords, driving up CPCs for both themselves and the Morgan Stanley corporate office. This posed a challenge as these regional financial advisors considered themselves to effectively be small business owners concerned with their own marketing and customer acquisition.
  • Collins solved this problem by unifying all SEM efforts through the home office, establishing a structured hierarchy that included national, regional and local campaigns that acted as “swim lanes” to keep financial advisors from accidentally bidding up on national campaign keywords.
  • By researching and building out a list of national, regional and local keywords and ads, the Morgan Stanley team was able to coordinate a tiered system of ads for which local queries for financial services would always return local advisors as the top query. The corporate team also built out a centralized repository of educational resources to answer frequently-asked questions, as well as case studies documenting specific successes – which, over time, more-quickly facilitated adoption from other financial advisors in the company’s network.

>>Back to top

Leveraging Partners to Manage the “Digital Shelf”
Featuring: @amykayaks

  • Shipman explains that in her capacity as digital strategist at HP, her primary goal is to promote sales of ink and toner cartridges – a process that has changed substantially due to the advent of online shopping. The customer journey of 10+ years ago involved customers first choosing physical storefronts to visit, then picking up a product off the shelf. This emphasized the importance of dominating shelf space (which may be why the toner aisle of your local store has row after row and shelf after shelf of various HP printer cartridges).
  • With online shopping, customers choose the product first, and pay significantly less attention to the merchant selling that product. As a result, marketing become a battle for customer attention online for what Shipman calls “the digital shelf” – or search engine results page (SERP). While her search teams were able to consistently hit #1 placement for paid and organic listings, she noticed that positions #2-5 would be filled with competitors, giving them sufficiently more page real estate than HP controlled in total.
  • Shipman describes 3 principles of maximizing shelf presence:
    1. Customer experience on SERP as a whole – thinking holistically about the full appearance  of SERP
    2. Focus on coverage of product, not URL rank – It’s arguably less important who sells HP ink, as long as it’s HP ink that’s sold
    3. Partners are critical for maximizing coverage
  • To maximize her digital shelf presence, Shipman used these 5 steps:
    1. Conduct keyword research – This involved comprehensive qualitative study, including hours of international interviews, as well as quantitative analysis of how many keywords would be necessary and how much would be required to buy them
    2. Create a click curve – Shipman describes this as her “hack” to dominate virtual shelf space by assigning values to every SERP position (not just position #1), and valuing real estate by channel.
    3. Enable search tracking – By tracking keywords, Shipman’s team was able to determine which keywords showed up where and how often for both paid and organic listings.
    4. Rate resellers – The biggest part of Shipman’s “hack,” this rating system gave her team an at-a-glance dashboard showing the position dominance of different types of rated resellers. Resellers that primarily sold HP products above-the-board were considered “friends”; resellers that strongly de-emphasized HP products were considered “foes.” This aggregate dashboard gave her team the ability to see that paid position #1 on SERP might primarily consist of traffic from “friendly” retailers, while top organic positions on SERP might primarily consist of traffic from “foe” retailers, which would require her SEO team to take action to strengthen “friendly” retailers’ online presence on SERP. Shipman’s team aggressively uses co-op partnerships for both paid and organic to ensure her teams can boost paid, organic and Google Shopping rankings for HP products. The HP team is also exploring using shared AdWords accounts for affiliates to provide even greater performance transparency.
    5. Project product coverage – Combining all these elements provides an aggregate, product-based view of the digital shelf.

>>Back to top

Co-Op Marketing Shared, Distributed and/or Centrally Controlled
Featuring: @kmullett

  • Co-op search marketing can be challenging when partners/franchisees are not on the same page as you are. It should be noted that neither domain authority nor backlinks are necessarily a “silver bullet” to guarantee strong rankings on SERP. Poorly-done websites will never perform, regardless.
  • This is why it’s worthwhile to consider looking into SERP structure, including name, address, phone (NAP) packs, local social accounts and PLA listings, which may vary wildly depending on query type. Even high-quality content may fail to perform if it lacks proper context or relevance to the searchers, their needs and their locations.
  • The key is efficient and affordable personalization of relevant content and delivery at scale, while allowing optional local participation. It’s potentially preferable to have a central content management system to ensure distributors can update their websites with up-to-date product info, rather than rely on occasionally unreliable franchisees who may not prioritizing updating their websites with the latest information.

>>Back to top of Day 2 session recaps

>>To SMX East Day 1 Recaps 

>>To SMX East Day 3 Recap